Chronic bacterial infections are inherently resistant to treatment. This is true even if organisms are antibiotic-sensitive, and high concentrations of drugs reach infection sites. The infections that afflict patients with the genetic disease cystic fibrosis (CF) are a prime example. Once infection is established it cannot be eradicated, and lung dysfunction caused by chronic infections claim the lives of the vast majority patients. Importantly, the mechanisms producing treatment resistance in CF and other chronic infections are poorly understood. Here we exploit the infrastructure of an ongoing clinical trial, and new findings about genetic diversity within infecting P. aeruginosa populations as tools to study treatment resistance. In preliminary studies, we found that infecting P. aeruginosa strains evolve to produce genetically diverse (but clonally-related) bacterial populations. We found that the relative abundance of subpopulations change as patients are treated with antibiotics. This is important because subpopulations that increase in abundance during treatment possess phenotypes that enable them to resist treatment in vivo, while subpopulations that decrease lack resistance functions. Studying these subpopulations could identify the mechanisms used by bacteria to withstand antibiotics in chronic human infections. Aim 1. Identify P. aeruginosa subpopulations and tolerance phenotypes linked to in vivo resistance. We will identify and bank variant subpopulations that show sensitivity and resistance to antibiotic treatment in vivo. We will use the sensitive and resistant subpopulations to test hypotheses that link treatment resistance to antibiotic tolerance mechanisms. We will also use an unbiased approach to identify phenotypic differences in sensitive and resistant subpopulations. Aim 2. How do sensitive and resistant subpopulations differ in protein expression? To identify candidate bacterial functions that may produce in vivo resistance, we will compare the proteomic profiles of sensitive and resistant subpopulations. This work will identify proteins and regulatory pathways associated with treatment resistance across multiple patients, a key first step in finding bacterial functions that mediate resistance in chronic infections. Aim 3. Which P. aeruginosa proteins likely treatment resistance in vivo? We will use several analyses to determine which differentially expressed P. aeruginosa proteins are most likely to contribute to treatment resistance in vivo. We will prioritize proteins exhibiting parallel changes in multiple patients, and those showing changes consistent with the tolerance phenotypes. We will then use engineered P. aeruginosa strains to link these proteins to tolerance mechanisms. Finally, we will use multiple reaction monitoring mass spectrometry (MRM) to determine which bacterial proteins show consistent expression changes in sputum from CF patients.